Some graduate students in science, technology, engineering and math—or STEM—only do research, under the guidance of a mentor. Other STEM grad students also have teaching responsibilities, for example, instructing undergrads or local high schoolers. Now a study finds that grad students who also teach show significant improvement in written [--1--], compared with grad students with no teaching requirement. The finding appears in the journal Science.

Researchers followed almost a hundred grad students attending three different institutions. About half of the students also taught. All the grad students were asked to submit written research proposals at the beginning of an [--2--] and a revised version again at the end. And those grad students who spent the year doing research and teaching showed bigger improvements in coming up with [--3--] hypotheses and in the design of valid experiments.

Differences in overall written quality among the students could not account for the results, because only [--4--] among those analyzed showed improvement as a function of the teaching experience. So teaching may make STEM grad students better scientists. Not to mention better teachers.
【视听版科学小组荣誉出品】
research proposals academic year testable specific skills
工科研究生的写作“秘籍” 有些科学、技术、工程和数学（简称STEM）的研究生仅在导师的指导下做一些研究。其他STEM的研究生则还有教学的责任，比如说指导本科生或者当地的高中生。最近，一项研究发现有教学责任的研究生与没有教学要求的研究生相比，在撰写研究报告时有显著提高。这项发现发表在《科学》期刊上。 研究人员跟进三所不同机构近100名研究称。其中有一半的学生有教学经验。所有研究生被要求在学年初交一份研究报告，在学年底递交修订版。而那些花了一年时间做研究和教学的学生在提出可检验性假设和有效实验设计中有很大提高。 学生总体撰写质量的差异并不能解释研究结果，因为在所有分析数据中仅仅这一项特定的技能显示出拥有教学经验的好处。因此教学可能会使STEM研究生成为更优秀的科学家，更不必说还会是更好的老师了。